Free-Space Detection with Self-Supervised and Online Trained Fully Convolutional Networks

Sanberg, Willem P.; Dubbelman, Gijs; With, Peter H. N. de

doi:10.48550/arxiv.1604.02316

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following a very similar idea, [17] recently proposed a method for free-space detection in urban environments with a self-supervised and online trained Fully Convolutional Network (FCN) [18]. In this case, the automatic ground truth collection process is based on a combination of disparity maps and Stixel-based ground masks computed from a calibrated stereo-pair.…”

Section: Related Workmentioning

confidence: 99%

Detecting unexpected obstacles for self-driving cars: Fusing deep learning and geometric modeling

Ramos

Gehrig

Pinggera³

et al. 2017

2017 IEEE Intelligent Vehicles Symposium (IV)

141

View full text Add to dashboard Cite

Abstract-The detection of small road hazards, such as lost cargo, is a vital capability for self-driving cars. We tackle this challenging and rarely addressed problem with a vision system that leverages appearance, contextual as well as geometric cues.To utilize the appearance and contextual cues, we propose a new deep learning-based obstacle detection framework. Here a variant of a fully convolutional network is used to predict a pixel-wise semantic labeling of (i) free-space, (ii) on-road unexpected obstacles, and (iii) background. The geometric cues are exploited using a state-of-the-art detection approach that predicts obstacles from stereo input images via model-based statistical hypothesis tests.We present a principled Bayesian framework to fuse the semantic and stereo-based detection results. The mid-level Stixel representation is used to describe obstacles in a flexible, compact and robust manner.We evaluate our new obstacle detection system on the Lost and Found dataset, which includes very challenging scenes with obstacles of only 5 cm height. Overall, we report a major improvement over the state-of-the-art, with relative performance gains of up to 50%. In particular, we achieve a detection rate of over 90% for distances of up to 50 m. Our system operates at 22 Hz on our self-driving platform.

show abstract

Section: Related Workmentioning

confidence: 99%

Detecting unexpected obstacles for self-driving cars: Fusing deep learning and geometric modeling

Ramos

Gehrig

Pinggera³

et al. 2017

2017 IEEE Intelligent Vehicles Symposium (IV)

141

View full text Add to dashboard Cite

show abstract

“…A substantial amount of research is devoted to road and lane detection [8] with free-space approximation using cameras, Lidar, or a fusion of both. Camera images can either be processed using classical methods [9] or machine learning techniques [10]. Na et al [11] show that even in complex road situations the drivable free-space can be extracted from Lidar point cloud data.…”

Section: Related Workmentioning

confidence: 99%

Roadgraph Generation and Free-Space Estimation in Unknown Structured Environments for Autonomous Vehicle Motion Planning

Keßler

Minnerup

Esterle

et al. 2018

2018 21st International Conference on Intelligent Transportation Systems (ITSC)

View full text Add to dashboard Cite

Automotive manufacturers and customers wish to have fully automated driving functionality available in a huge set of locations, scenarios, and markets. This raises the need for universally applicable scene understanding and motion planning algorithms that do not rely on highly accurate maps or excessive infrastructure communication. In this paper we introduce two novel approaches for extracting a topological roadgraph with possible intersection options from sensor data along with a geometric representation of the available maneuvering space. Also, a search and optimization-based path planning method for guiding the vehicle along a selected track in the roadgraph and within the free-space is presented. We compare the methods presented in simulation and show results of a test drive with a research vehicle. Our evaluations show the applicability in low speed maneuvering scenarios and the stability of the algorithms even for low quality input data.

show abstract

Free-Space Detection with Self-Supervised and Online Trained Fully Convolutional Networks

Cited by 2 publications

References 0 publications

Detecting unexpected obstacles for self-driving cars: Fusing deep learning and geometric modeling

Detecting unexpected obstacles for self-driving cars: Fusing deep learning and geometric modeling

Roadgraph Generation and Free-Space Estimation in Unknown Structured Environments for Autonomous Vehicle Motion Planning

Contact Info

Product

Resources

About